IDA-MCP

IDA-MCP

An MCP server that runs inside IDA Pro for AI tools, exposing reverse-engineering capabilities like function listing, decompilation, and analysis via HTTP.

Category
Visit Server

README

IDA-MCP

MCP server that runs inside IDA Pro 9.4 so tools like Cursor / Claude / Cline can talk to your open IDB over HTTP.

Default listen address: 127.0.0.1:8765
Stack: IDAPython + Hex-Rays (when present). No extra pip deps inside IDA.

Install

.\install.ps1 -IdaPath "C:\Program Files\IDA Professional 9.4"

Copies the plugin to %APPDATA%\Hex-Rays\IDA Pro\plugins\.

Uninstall:

.\uninstall.ps1

Usage

  1. Open IDA 9.4 and load an IDB
  2. Edit → Plugins → IDA-MCP → Start (or Ctrl-Alt-M to toggle)
    • Stop: Edit → Plugins → IDA-MCP → Stop (or Ctrl-Alt-M again)
  3. Point your MCP client at http://127.0.0.1:8765/mcp

Cursor

%USERPROFILE%\.cursor\mcp.json or project .cursor/mcp.json:

{
  "mcpServers": {
    "ida-mcp": {
      "url": "http://127.0.0.1:8765/mcp"
    }
  }
}

If you set a token in config:

{
  "mcpServers": {
    "ida-mcp": {
      "url": "http://127.0.0.1:8765/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_TOKEN"
      }
    }
  }
}

More examples under examples/.

Smoke check

Invoke-RestMethod http://127.0.0.1:8765/health

Try: call get_database_info, then list_functions / decompile_function.

Config

%LOCALAPPDATA%\IDA-MCP\config.json

{
  "host": "127.0.0.1",
  "port": 8765,
  "token": "",
  "allow_write": false,
  "allow_patch": false,
  "allow_execution": false,
  "max_read_size": 1048576,
  "max_call_depth": 20,
  "log_level": "INFO"
}

Env vars override the file: IDA_MCP_HOST, IDA_MCP_PORT, IDA_MCP_TOKEN, IDA_MCP_ALLOW_WRITE, IDA_MCP_ALLOW_PATCH, IDA_MCP_ALLOW_EXECUTION, …

Logs: %LOCALAPPDATA%\IDA-MCP\ida-mcp.log

Defaults

Bind localhost only
Rename / comments / types off (allow_write)
Byte patching off (allow_patch)
execute_idapython off (also needs a token)
Memory reads capped at 1 MiB

Don't bind 0.0.0.0 without a token — the server refuses that.

How it works

HTTP JSON-RPC (Streamable HTTP) in a background thread inside IDA. Tool handlers schedule IDA API calls with ida_kernwin.execute_sync so the UI doesn't get trashed from worker threads.

Stdlib only in-process; the optional mcp package in requirements.txt is for external tooling/tests, not required by the plugin.

See docs/architecture.md.

Tools (high level)

There are ~57 tools. The useful ones day-to-day:

  • DB: get_database_info, get_entry_points
  • Functions: list_functions, get_function*, get_disassembly, decompile_function
  • Xrefs / graph: get_xrefs_*, get_callers, get_callees, trace_call_path
  • Imports / strings: list_imports, find_import, list_strings, find_string
  • Search / memory: search_bytes, search_text, read_memory
  • Analysis helpers: analyze_function, summarize_function

Writes and patches are separate and gated. Full list: docs/tools.md.

Resources: ida://database, ida://function/0x…, …/pseudocode, …/disassembly.

After code changes

Either restart IDA, or run scripts/reload_ida_mcp.py via File → Script file.

Tests

pip install pytest
pytest tests -q

Live checks (IDA must be running with MCP up): tests/_live_mcp_check.py, tests/_live_full_test.py.

Layout

ida_plugin/          plugin entry + package
install.ps1          Windows installer
docs/                architecture / tools / install notes
examples/            MCP client configs
.cursor/skills/      Cursor skill for using the tools
tests/

License

MIT — see LICENSE.

Recommended Servers

playwright-mcp

playwright-mcp

A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.

Official
Featured
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.

Official
Featured
Local
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.

Official
Featured
Local
TypeScript
VeyraX MCP

VeyraX MCP

Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.

Official
Featured
Local
graphlit-mcp-server

graphlit-mcp-server

The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.

Official
Featured
TypeScript
Kagi MCP Server

Kagi MCP Server

An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.

Official
Featured
Python
E2B

E2B

Using MCP to run code via e2b.

Official
Featured
Neon Database

Neon Database

MCP server for interacting with Neon Management API and databases

Official
Featured
Exa Search

Exa Search

A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.

Official
Featured
Qdrant Server

Qdrant Server

This repository is an example of how to create a MCP server for Qdrant, a vector search engine.

Official
Featured